How Eon unified multi-cloud incident tooling behind one MCP Gateway on AgentCore
Eon, a cloud data protection vendor, replaced per-tool credentials and local MCP servers with a single authenticated MCP gateway on Amazon Bedrock AgentCore. Developers and an autonomous SRE agent now investigate production incidents across AWS, GCP and Azure through one URL with policy-based access.
The problem
Eon's engineers investigate production incidents across several clouds and dozens of operational systems, so one issue can involve logs, metrics, pull requests and cloud health checks. Each tool needed its own credentials, and each developer ran several local MCP servers. New engineers spent about two days on setup, and the SRE agent needed the same tools as humans but with narrower permissions.
How they did it
- Deploy AgentCore Gateway as one company-wide MCP endpoint, with Policy, Runtime and Identity alongside it.
- Federate corporate SSO through Amazon Cognito to issue JWTs for humans, and use client_credentials grants for agents.
- Host MCP servers on AgentCore Runtimes across several AWS accounts, converting local stdio servers into Streamable HTTP endpoints, and let the gateway reach them through IAM cross-account role assumption.
- Write Cedar policies in ENFORCE mode mapping SSO groups and agent identities to permitted tools, kept as code and reviewed through GitHub PRs.
- Use DYNAMIC listing mode for large tool catalogs to keep policy size down, and wrap some remote MCP servers in a Runtime that filters their tool lists.
- Add an Amazon API Gateway shim so Claude Code, Cursor and Codex can complete the OAuth flow.
Results
- As reported by AWS (guest post co-written with Eon): developer onboarding dropped from 2 days to 2 minutes.
- As reported by AWS: local MCP processes per machine went from 10 to 0, and per-system credential sets were replaced by one SSO login.
- As reported by AWS: the setup serves 100+ developers and about 10,000 sessions, with hundreds of tools across dozens of targets behind one URL.
- VPC peering between accounts was replaced by IAM trust policies, and one Cedar policy set now governs both humans and agents.
As reported by the source (AWS Machine Learning Blog (guest post)); AgentGid did not measure these figures.
This suits platform or SRE teams already running on AWS, comfortable with Cognito/SSO federation, IAM cross-account roles and Cedar policy writing. Treat the 2-day-to-2-minute onboarding figure as AWS-reported, and expect the policy and OAuth shim work up front. If you only need the coding client side, OpenCode (Free + $10/mo, open source) is a cheaper alternative to Claude Code ($20/mo).
The agents used here
Anthropic's agentic coding tool that reads a codebase, edits files and runs commands from the terminal, with VS Code and JetBrains integrations plus …
- Price
- $20/mo
- Gid Score
- 85 #1 in Terminal
AI coding editor and agent platform with a desktop app, a CLI, web and mobile agents, and cloud agents that run tasks in remote environments. It offe…
- Price
- Free + $8/mo
- Gid Score
- 71 #1 in IDEs
OpenAI's coding agent, offered as an open-source terminal CLI, an IDE extension, a desktop app and a cloud agent that works on repositories. Users si…
- Price
- Free + $8/mo
- Gid Score
- 84 #2 in Terminal
Compare: Claude Code vs Cursor · Claude Code vs OpenAI Codex · Cursor vs OpenAI Codex
Similar use cases
Mod almost any PC game with Claude Code and the universal-modder plugin
An open-source GitHub project packages skills, a command-line tool, a shared knowledge base and the fal MCP server so AI coding agents such as Claude Code can mod PC games the use…
The repository ships three worked examples: a Terraria tModLoader weapon and boss pack with sprites made through fal, an Age of Empires II DE civiliz…
Let Claude Code and Codex point at Mac screen elements with bigarrow
An open-source, MIT-licensed macOS command-line tool called bigarrow, with a skill for Claude Code and Codex, lets AI agents draw a click-through arrow and text sign over the scre…
Drawing needs no macOS permission; only some target-finding modes need Accessibility or Screen Recording, granted to the terminal app that launched b…
Route between coding-agent models with the open-source Weave Router 2.0
The Weave team presented Weave Router 2.0 in a Show HN post. It is an open-source routing model that sits in front of coding agents and switches between LLMs per turn, and the aut…
As reported by the Weave author on Hacker News: on Terminal Bench 4.0 and SWE Atlas, the router had equivalent pass rates to GPT-6 Astra.
Guides
Sonnet Builds, Opus Reviews, Haiku Reads: Does the Claude Code Advisor Save Money?
A viral post says running Sonnet, Haiku and Opus together in Claude Code gives better code for fewer tokens. We checked it against Anthropic's docs and prices: when the advisor pays off, when it costs more, and why a ready project kit matters more than the model mix.
20 Claude Code Plugins 'Anthropic Just Dropped': Not From Anthropic, Checked
A viral post says Anthropic dropped 20 Claude Code plugins worth $29,077. We checked all 20 GitHub repos: community projects, none in Anthropic's directory, the price tag is the post’s own arithmetic, one is not open source, and the stars mostly hold.
Videos Made in Code With Opus 5.5: 8 Popular GitHub Repos, Checked
Prompt collections, two P(doom) music videos, a style pack, a cartoon starter kit and a Claude Code skill: we checked the 8 repos people share for Opus 5.5 code-made videos, their licenses, what they need to render and what you can reuse.
Claude Code Deploys: 6 Open-Source Repos for a Containment System, Checked
A viral post says GitHub Actions, Argo CD, Argo Rollouts, Prometheus, Sentry and OpenTelemetry turn Claude Code deploys into a containment system. We checked each: role, license, what needs Kubernetes, and what Claude Code's own permissions add.